I’ve been mulling over one of Kant’s arguments on duties owed to animals. The ethical system Kant lays out is centered on mutual respect between two humans, on the basis of recognizing the others as rational creatures like ourselves. He makes the argument at some point that who we are failing when we are cruel to animals is ourselves, because making ourselves comfortable with cruelty harms our sense of empathy and hurts our relationships with other humans.
If a man shoots his dog because the animal is no longer capable of service, he does not fail in his duty to the dog, for the dog cannot judge, but his act is inhuman and damages in himself that humanity which it is his duty to show towards mankind. If he is not to stifle his human feelings, he must practice kindness towards animals, for he who is cruel to animals becomes hard also in his dealings with men. (Per SEP:
[1784–5] 1997: 212 [Ak 27: 459])
This is an argument that is a little bit unsatisfying to a lot of modern readers. In the couple hundred years since Kant’s time our views on animal welfare have slowly shifted towards a higher regard for other animals. Most of us think that being cruel to an animal is wrong because it is bad for the animal, whether or not our cruelty harms ourselves in some way.
But while other authors have put forward more satisfying arguments for animal welfare from a Kantian perspective (I particularly enjoyed Fellow Creatures), I’ve been coming back to this argument in the context of chatbots. As more people get used to using chatbots, especially ones that are good at simulating sensitivity to subtle language cues, we see funny articles about people treating machines with human formality. Major news outlets like the USA Today were running stories on the costs associated with saying “thank you” to a chatbot.
I think in conversations with a chatbot we get to see a more pure version of Kant’s argument for exercising humanity even when dealing with nonhuman creatures. There are some fringe views but I think the majority judgement of LLMs is that what they are is not sufficient to really be called a mind. It’s good at simulating a conversation but it doesn’t seem like it has an experience of the world and the passage of time. Thus, most of us don’t think that being rude to a chatbot is bad for the chatbot, because we don’t think it is the kind of thing that is capable of experiencing injury. It can give you back a response that sounds like what someone who has been treated rudely, but it can’t actually feel shame or distress at the loss of social standing.
Kant argues that we still have a duty to demonstrate humanity, because when we grow accustomed to acting in inhumane ways we damage our own humanity. This is especially relevant given that the chatbot is very good at giving the impression of having a conversation with another human. Chatbots increasingly interact with us in the voice of a coworker, a friend, or a neighbor. Practicing meeting that voice with patience and courtesy seems worthwhile even if there’s nobody on the receiving end.